Comparing the accuracy of unmanned aerial vehicle based three-dimensional reconstruction methods for large-scale onion fields

— Photogrammetry vs. light detection and ranging —

Abstract

Three-dimensional (3D) information on crops and soil undulations is important for precision agriculture. Light detection and ranging (LiDAR) has previously been used to obtain highly accurate 3D information. However, with recent advancements in digital technology, photogrammetry, which does not require special sensors, has attracted considerable attention. In this study, models using these two 3D reconstruction methods are compared and their corresponding applications are summarized. The results show that the standard deviation of the height difference between the 3D models of the two methods is 4.1 cm in a 1.7 ha onion field. However, it is inferred that LiDAR is more suitable than photogrammetry for understanding the 3D growth information of slender upright onion plants.